This repository temporarily contains data from our study on the causes for misclassifications of Steam game reviews.
The data consists of a sample of 100K URLs of game reviews from Steam with the associated sentiment and it is available in three formats:
- CSV
- XLSX
- TXT
The sentiment of each review was computed by three different sentiment classifiers:
- NLTK
- SentiStrength
- Stanford CoreNLP